Meta has unveiled a slew of enhanced protective measures against child sexual exploitation on Facebook and Instagram, including autonomous red-teaming AI agents and Large Language Model (LLM) detection tools. The goal is to break up sophisticated predator networks and target elusive advertising schemes that try to direct users to illegal material hosted outside of its platforms.
After a policy update in September, Meta now reports child safety offenses directly to the National Cyber Crime Reporting Portal (cybercrime.gov.in) under India’s Indian Cybercrime Coordination Center (I4C). This direct pipeline adds to its existing protocol of working with the domestic authorities in conjunction with help from NCMEC.
Meta said it took action on 33.2 million pieces of child sexual exploitation (CSE) material globally between January and June 2026, with more than 97% proactively detected by the company before being reported by users. During the same six-month period, the company alone took down 5.3 million pieces of violating content in India, more than 98% of which was proactively flagged and resolved.
To fight these stealth marketing funnels, Meta introduced several new automated tools to strengthen its ad review pipeline: LLM “Signposting” Detection: New language models scan seemingly innocuous ad copy to detect subtle linguistic cues and hidden hints directing users to illegal off-platform destinations. Online predators have increasingly turned to paid promotional tools to create ads that look innocuous on the surface but are actually directing people toward illegal, off-platform child sexual abuse material (CSAM). Deep Landing Page Analysis: Systems now look at where an ad leads, not just the promotional asset itself, meaning Meta can block malicious destination websites and penalize the ad accounts responsible. Red-Team AI Agent – an internal AI agent that continuously tests and attacks Meta’s own defenses to uncover security blind spots, to discover new adversarial tactics before bad actors can deploy them at scale. Meta takes down posts when it finds violating material, disables bad actor accounts, and reports findings to the National Center for Missing & Exploited Children (NCMEC) as required by law.